Solodov, M. Convergence analysis of perturbed feasible descent methods. Journal of Optimization Theory and Applications (93)2:337--353, 1997. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Reinforcement Learning Through Gradient Descent - Baird, III (1999) (7 citations) (Correct)
....based on the general literature for stochastic approximation. The convergence of stochastic and incremental algorithms for neural networks has been extensively studied (White 1989, White 1990, Gaivoronski 1994, Mangasarian Solodov 1994, Luo Tseng 1994, Solodov 1995, Mangasarian Solodov 1995, Solodov 1996, Luo 1991, Bertsekas 1995, Bertsekas Tsitsiklis 1996, Solodov 1997, Solodov and Zavriev 1998) Over the last few years, results have been extended and generalized. Two of the latest papers are most relevant to the algorithms in this thesis. If the f(x) function is smooth and has a Lipshitz ....
Solodov, M. V. (1996). Convergence Analysis of Perturbed Feasible Descent Methods.
On Approximations With Finite Precision In Bundle Methods For.. - Solodov (2002) Self-citation (Solodov) (Correct)
No context found.
M.V. Solodov. Convergence analysis of perturbed feasible descent methods. Journal of Optimization Theory and Applications, 93:337-353, 1997. 14
Nonmonotone And Perturbed Optimization - Solodov (1995) (2 citations) Self-citation (Solodov) (Correct)
.... number of the partial error functions evaluations) However, for some problems this is still practical [23] 61 Chapter 3 Convergence Analysis of Perturbed Feasible Descent Methods We develop a general approach to convergence analysis of feasible descent methods in the presence of perturbations [63]. The important novel feature of our analysis is that perturbations need not tend to zero in the limit. In that case, standard convergence analysis techniques are not applicable. Therefore a new approach is needed. We show that, in the presence of perturbations, a certain approximate solution ....
M.V. Solodov. Convergence analysis of perturbed feasible descent methods. Journal of Optimization Theory and Applications, June 1995. Submitted.
Incremental Gradient Algorithms With Stepsizes Bounded Away From.. - Solodov (1998) (3 citations) Self-citation (Solodov) (Correct)
....We point out that Theorem 2.2 is the first convergence result of any kind for incremental algorithms with stepsizes bounded away from zero. Our analysis is by virtue of characterizing IGA as a special perturbed gradient method (Proposition 2. 1) and it makes use of some of the ideas employed in [19] where general perturbed feasible descent algorithms are studied. It is worth to point out that the main results of this paper cannot be obtained using the approach of [13] As an aside, we show that under a certain additional assumption on the growth property on the gradients (similar to the one ....
....then every accumulation point x of the sequence fx i g has the above property. Proof. The result follows from combining Proposition 2.1 and Theorem 2.1. The analysis presented here can be applied to a variety of modifications and extensions of Algorithm 1.1. For example, using the approach of [19], we could treat the projection version of IGA described in [10, 21] At the expense of introducing considerably more notation and some technical details, we could also consider the parallel and momentum term modifications given in [13, 21] as well as algorithms with noisy data along the lines of ....
M.V. Solodov. Convergence analysis of perturbed feasible descent methods. Journal of Optimization Theory and Applications, 93: 337--353, 1997.
Unknown - World Congresses Of (Correct)
No context found.
Solodov, M. Convergence analysis of perturbed feasible descent methods. Journal of Optimization Theory and Applications (93)2:337--353, 1997.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC